01 · Learn
Move from CPU, memory, and I/O to thread pools, event loops, and coroutines in a causal sequence.
Free 16-lesson courseFrom the instant a CPU executes one instruction to the moment Spring, a database, or Kafka becomes the bottleneck. Learn the mechanism, then reproduce it yourself.
The public course starts without sign-up.
SIMULATION EXAMPLE
The bottleneck is one serialized inventory row, not the app server.
Open this scenario and run an experiment →ONE LEARNING LOOP
The course and simulator are one system. Revisit each mechanism as load, queues, pools, and latency — not as a disconnected diagram.
Move from CPU, memory, and I/O to thread pools, event loops, and coroutines in a causal sequence.
Free 16-lesson courseChange RPS and pool sizes. Wire caches, brokers, and replicas. Watch queues fill and requests drop.
Run the simulatorPractice cause, impact, and trade-offs using evidence such as CPU 96%, queue 400, and p99 1.8s.
Browse scenariosFREE + PRO
Simulation accuracy is not a paywall. Pro focuses on the work that carries ongoing creation and operating cost: deep lessons, cloud sync, and AI.
Learn, then verify it yourself
The core engine and public learning path stay open.
Enough to learn how servers work and test your own hypothesis with real scenarios.
Follow the mechanism all the way down
Deep lessons, AI feedback, and a workspace that follows you.
For backend engineers who want to explain the cause, not just remember the conclusion.
BUILT FOR EXPLANATION
Watch queues, pools, and latency change with simulator state. The benchmark page explains throughput calibration and the limits of p99 and production predictions.
Experiment with threads.max, pool size, accept-count, heap, GC, and replica lag — the values used during real incidents.
Save before and after a cache or scale-out change, then compare success rate, p99, and drops side by side.
START SMALL
No prerequisites, 5–10 minutes at a time. Once it clicks, find the same mechanism in the simulator.